Implement timing literals as literals with suffix

[jlapeyre]

The spec currently sort of implies that there must not be whitespace between the number and suffix of a timing literal. That is

Good

100ns;
1.02s;

Bad

100 ns;
1.02 s;

However this parser currently allows spaces. The reference parser does as well.

It would simplify the parser (wide sense) if spaces were not allowed. Alternatively, we could make all of "100 ns" a token in the lexer. I'm pretty sure that there are no other tokens that are handled specially in this way at the moment.

Best would be to disallow spaces. But perhaps we should check for clarification with the TSC.

[jwoehr]

Was this clarified?